Abstract: Beyond Foundation Models: A Sigmoid-Based Hybrid Transition for Distilling DINOv3 Features into Residual Encoders for Medical Segmentation

<iframe src="https://www.googletagmanager.com/ns.html?id=GTM-MQDXN445" height="0" width="0" style="display:none;visibility:hidden"></iframe>

Abstracts

AI in Medicine 2026

Beyond Foundation Models: A Sigmoid-Based Hybrid Transition for Distilling DINOv3 Features into Residual Encoders for Medical Segmentation

by Bahareh Etaati

Bahareh Etaati